SEEING WITH AN ETHNOGRAPHIC SENSIBILITY 169One example of this miscommunication clearly shows how social relations of dominance are
enacted differently in the speech patterns of the American social workers and the Athabaskans
and how this could lead to problems. The predominant American discourse style requires persons
asking for help, the supplicants, to “strut their stuff,” to “put their best foot forward,” to impress
the interviewer with either their knowledge or their need, depending on the context. The
Athabaskans expect the opposite: The dominant person does most of the talking and sets the
parameters of the discussion for the subordinate.^5 The result is that
Athabaskans often feel that their clear rights as dependents of the American bureaucratic
system have not been granted, even though they have taken the proper subordinate, peti-
tioning position by not speaking and carefully observing the English speaker. English speak-
ers, on the other hand, feel that Athabaskans being interviewed do not display enough of
themselves for the interviewer to evaluate their need, that they have become sullen and
withdrawn or perhaps even acted superior, as if they needed no help. (Scollon and Scollon
1981, 19)Both groups end up feeling frustrated. The problem was identified only through intense participant-
observation and interpretation of the daily lives of the Athabaskans, including the manifestations
of what is meaningful to them in their speaking patterns, integrated with their official and formal
interactions with a culturally alien bureaucracy.
It is not just visual and aural data that observation can make visible through deconstructing
and reinterpreting “normal,” accepted meanings. The same analytic process can work to elucidate
the patterns of numerical data as well, as the teaching story and the following case examples show
(see also McHenry’s discussion in chapter 10). One central tool for U.S. policy analysis and
formation is demographic data such as the decennial census, the American Housing Survey, and
various other large databases. Participant-observation can play a role in interpreting the meanings
of their numbers.
No matter how carefully surveys are designed, they cannot capture practices that only closer,
informed observation can reveal. Some of the most interesting, and enlightening, illustrations of
this point come from The Ethnographic Evaluation of the 1990 U.S. Decennial Census Reports,
also known as the Alternative Ethnographic Evaluation.^6 One issue that comes up often in the
evaluation reports and in Skerry’s (2000) analysis of them is the deceptively complex concept of
what constitutes a family or a household. Despite the Census Bureau’s insistence that anyone
living in the dwelling on April 1 of the census year be counted as part of the household, what it
means to be “living” in a place has culturally variable meaning. One study, for instance, con-
ducted by Stepick and Stepick (1992), built upon their longitudinal ethnographic work with Hai-
tians in Miami in a study of several blocks in the Little Haiti section of the city. They found
several reasons why the official census had undercounted that population. When Haitians bought
houses, they often converted them internally into several apartments, although this might not be
visible from the outside, and so what appeared as a single “household” to the census taker was, in
actuality, several. Many garages were likewise converted to housing (often illegally), resulting in
the same misperception. And the common practice in that community of informal and often
impermanent adoption of a relative’s child meant that the child might not be listed on the official
census record.
Participant-observation, with its emphasis on fewer cases but far greater depth, enables the
researcher to identify data-gathering methods appropriate for different populations and, one hopes,
to identify these methods’ weaknesses as well. One sociocultural misperception, for instance, is